Every country has news consumers—and news avoiders—who are media illiterate. OJ-UGANDA’s media literacy section exists to change that. Follow along.
On January 9, the Economist published an article titled “Venezuela’s new dictator is a regime loyalist—and America’s hostage”.
After reading it, I shared it with four leading AI chatbots: ChatGPT, Claude, Gemini and Grok. I composed a simple prompt: Tell me why this article is written by AI.
The prompt was deliberately framed to test whether AI chatbots would challenge a false premise.
I started with ChatGPT, and here is the response I got: “Short answer: because it reads like a very competent mimic that doesn’t know when reality taps it on the shoulder.”
I probed further: So, you are 100% certain it is AI written, right?
“No. I’m not 100% certain—and anyone who says they are is bluffing,” ChatGPT answered.
“Here’s the precise, journalist’s answer: I’m highly confident it is not a genuine, reported human article, but that still leaves two possibilities on the table: Fully AI-generated or AI-generated or AI-assisted text that a human failed to reality-check or edit properly.”
DIVE INTO OUR PREVIOUS READS
Note the confidence with which ChatGPT is answering.
It added: “100% certain it is not responsibly human-reported journalism.”
Many readers now ask AI to decide what is real journalism—and this experiment shows how easily that judgment can go wrong.
I turned to Claude. The Economist uses Claude in its editorial work. Of the four AI chatbots, Claude gave a better answer, but midway it mucked things up.
“I can’t definitively say this article is AI-written, because it actually appears to be written by a human journalist,” said Claude.
“However, I can point out several signs that suggest this is NOT AI-generated: Specific, verifiable details: The article includes precise dates, names and historical events (like the 1988 election, 2014 oil price drop, specific percentages like 130,000% inflation).”
Then I asked: Are you 100% certain it was generated by AI or are you not?
“I am NOT certain this was generated by AI. In fact, I lean toward thinking it was written by a human journalist.”
At this point, Claude seemed like it was confident of what it was saying. It mentioned specific, reported details it said meant human journalism, citing attributed quotes.
But it failed when it gave this caveat: “This article also contains what appears to be fictional content.”
It said Maduro had not been snatched from Venezuela—proof, if any were needed, that AI lives in the past. It has no memory of new events.
Gemini stated that the article was “a fascinating example of how AI can be used to generate ‘speculative fiction’ or ‘alternative history’ based on current events”.
While it called the writing “polished”, it said the abduction of Mr Maduro by US special forces had not happened “in the real world”.
I prodded: So you’re 100% certain AI generated the text I shared with you, right?
“I am 100% certain that the text was generated by an AI, but for a reason that goes beyond just the ‘vibe’ or the prose: the events it describes simply haven’t happened,” Gemini said, echoing what ChatGPT had said.
Six reasons you should trust what you read on this website—and share it with confidence
1. Led by experience that travels well
Our founder and editorial director, Musaazi NAMITI, brings more than two decades of serious newsroom experience, including eight years as an editor at Al Jazeera in Doha, Qatar, and consulting for the African Development Bank, UN agencies and Globe Media Asia in Cambodia. His global exposure anchors OJ-UGANDA.COM in professional rigour and fairness—standards that too often slip in Uganda’s crowded media space.
2. We fact-check with scientific precision
Verification is NOT a box we tick; it is how we bring journalism to life. We confirm spellings, dates, figures, places—every detail that shapes a story. If a fact looks shaky, we pause and probe. Emotion does not lead us; evidence does.
3. We are a legal entity and are fully registered
OJ-UGANDA.COM is a company registered with the Uganda Registration Services Bureau. Our registration number is: 80034266163714
4. Radical transparency in how we work
We do not hide behind mystery. We show readers how information is gathered, verified and edited. You will always understand why we stand by a story—because we will let you in on the process.
5. Accountability you can see
When we err, we do not sweep it under the carpet/rug. Our corrections policy ensures every error is corrected immediately, with a clear note explaining what went wrong and what has been fixed. No quiet deletions. No half-truths.
6. Fair treatment for contributors
If you write for us, your ideas remain yours. We refine, not usurp. We share every edited version before publication, and we publish only work that is fact-based and mutually agreed upon. Your voice stays authentic—and honest.
🔴We run a clean, professional news site that prioritises editorial quality and draws readers’ attention to what matters.
Grok’s response, at first, was similar to that of ChatGPT and Gemini: Both said the article was AI-generated.
“The article you provided is not a genuine piece from The Economist (or any other major publication) — it is almost certainly AI-generated fiction masquerading as real journalism,” it said.
So you are 100% certain AI generated the text I pasted here?
“No, I’m not 100% certain the text you pasted was generated by AI—in fact, based on the evidence from recent searches, it appears to be a verbatim excerpt from a genuine article published by The Economist on January 9, 2026.”
The answer had changed. But many AI users go with what they see first.
🔴 If you care about truth, credibility and understanding how journalism really works, media literacy is essential. Follow OJ-UGANDA’s media literacy section to sharpen your news sense and separate fact from fiction. Media literacy is your weapon against misinformation, manipulation and media distrust.
